As shown in FIG. 1, a hard disk drive 10 conventionally includes at least one rotating data storage disk 11 having its two major surfaces coated with a suitable magnetic coating (media). Frequently, but not necessarily, a plurality of disks 11 are mounted on a common rotating spindle to which rotational force is imparted by a suitable spin motor 13. Each data storage surface is provided with an associated head-slider 12 which "flies" in close proximity to the rotating data storage surface in accordance with so-called "Winchester" technology so as to write data to the surface, and to read data previously written to the surface. Each head slider is connected by a gimbal 14 to a load beam 16 which acts as a spring for applying a loading force (called "gram load") to the slider to urge it towards its associated disk surface. Each load beam is connected to a head arm 21 via a swage plate 18, which in mm is a part of an E-block 22 forming a rotary actuator. Each slider, gimbal, load beam and swage plate forms an assembly known as a "head-gimbal assembly" or "HGA".
A voice coil 24 converts electrical current into rotational force in combination with a magnetic field provided by e.g. two intense-field-providing permanent magnets 26. A read write channel 28 connects to a data transducer head formed on and carried by each slider. The read-write channel 28 is connected to disk drive electronics 30 which includes a head position servo loop 32 for controlling currents passing through the voice coil 24, and a spindle motor control 34 for controlling operation of the spin motor 14. A data path 36 leads to a host computer (not shown) with which the disk drive 10 is operatively associated. Data blocks are typically written in concentric data tracks defined on each data storage surface, and the head positioner actuator 22 moves the vertical stack of head-sliders 12 to each desired track location for writing and reading operations.
As shown in FIGS. 2A, 2B and 2C, each HGA as used in the hard disk drive 10 of FIG. 1 comprises four components: the slider 12, a gimbal 14, the load beam 16, and a swage or base plate 18; see e.g. FIGS. 2A, 2B and 2C. The load beam in the FIG. 2A example is a Hutchinson Technology type-8 suspension and a 70% slider is mounted to the load beam at the gimbal thereby to form the exemplary HGA.
The first component is the slider 12 which features a self-acting hydrodynamic air bearing and carries an electromagnetic transducer (hereafter "read/write head" or "head") which is used for recording and retrieving information from the spinning magnetic data storage disk. The head may be a thin film inductive read/write head or it may be a thin film inductive write, magnetoresistive read head structure. Other head designs, such as Metal-in-Gap (MiG) heads have also been widely used in flying head ("Winchester") magnetic recording technology.
The gimbal 14 is fixedly attached to the load beam 16 and is affixed to the slider 12 by a suitable adhesive, but is designed to be resilient in the slider's pitch and roll directions to enable the slider 12 to follow precisely the topography of the disk 11 over which it is flying. Also, the gimbal 14 is rigid in the yaw and in-plane directions for maintaining precise in-plane slider positioning.
The load beam 16 is typically formed of thin stainless steel alloy sheet material, which forms a base for attachment of the gimbal 14. The load beam 16 also defines the swage or base plate 18 which may be a sandwich construction of a thicker plate which is typically spot-welded to the load beam and which defines an annular swage flange 20 which is in turn attached by conventional ball swaging to an associated head arm 21 (see FIGS. 4A and 4B) of the actuator E-block 22. The load beam 16 further includes two sections: a resilient spring section 26 and a relatively rigid section 42 having e.g. longitudinal side flanges 44.
In the present example, the spring section 26 of the load beam 16 includes two longitudinal legs or web portions 26A and 26B. The spring section 26 is resilient in the vertical direction to permit the slider 12 to follow the topography of its adjacent data storage surface (see FIGS. 1 and 3A and 3B). The spring section 26 also supplies a downward force that counteracts the hydrodynamic lifting force developed by the slider in reaction to air pressure at the surface of the rotating data storage disk 11. In order to produce this counteracting force, the spring section 26 is plastically deformed into a curved shape so that applies a preload force to urge the gimbal 14 and slider 12 towards the disk surface. The spring preload force that counteracts the slider hydrodynamic force is commonly referred to in the art as "gram load".
The flanged section 42 is conventionally characterized by two vertical lip or flange portions 44 bent directly from the main body of the load beam providing additional, bending and stiffness to the load beam 16. The flange design could include up-facing vertical lip portions 44A, as shown in FIG. 3A for example, or the flange section design could employ down-facing vertical lip portions 44B as shown in FIG. 3B (this arrangement is usually known as an inverted-flange load beam design). One exemplary inverted-flange load beam design is disclosed and characterized in commonly assigned U.S. Pat. No. 5,027,241 to Hatch et al, entitled: "Data Head Load Beam for Height Compacted, Low Power Fixed Head and Disk Assembly", the disclosure thereof being incorporated herein by reference.
The fourth component is the base plate or swage plate 18 which attaches the load beam 16 to the support arm 21 of the actuator E-block 22. The base plate 18, also typically formed of stainless steel alloy, is often formed to have a greater thickness than the load beam so as to provide necessary stiffness for structural strength reason and to accommodate other features for fastening, such as the swaging flange or boss 20 which facilitates swaging to the head arm 21 of the actuator E-block, as shown in FIGS. 4A and 4C, for example.
An ideal head suspension design is one in which the slider faithfully and precisely follows the support arm motion within the frequency range of interest. That is, the E-block support arm motion generated by the actuator voice coil should be transmitted through the load beam and gimbal to the slider without any amplification, reduction, delay or lead, or any other distortion. When these conditions are obtained, the head tracking effort can be performed predictably and efficiently; and, track densities can thereby be increased and access times reduced. However, this result is difficult to achieve because any mechanical system that possesses mass and stiffness has resonant frequencies at which the input motion will be amplified substantially. FIG. 6 shows how the amplification varies with frequency of a prior art suspension design with up-facing flange (which in this example is a type-8 suspension manufactured by Hutchinson Technology, Incorporated, herein "type-8 HGA") from the support arm to slider. As shown in FIG. 6 for a type-8 HGA, the suspension support motion is amplified substantially at three separate resonant frequencies: the 1st torsional mode resonance (at approximately 2200 Hz), the sway mode resonance (at approximately 6000 Hz), and 2nd torsional mode resonance (at approximately 8500 Hz).
FIGS. 7A, 7B and 7C illustrate conventional finite element analysis models of the type-8 HGA. In this model of the suspension, three families of dynamic: modes are illustrated: namely, bending modes (FIG. 7A), torsional modes (FIG. 7B) and sway modes (FIG. 7C). Bending modes, when excited, cause the slider to move up and down relative to the magnetic disk, contributing slightly to a second order effect on the ability of the data head to follow the data track, if there is any discernible effect at all.
However, the torsional and sway modes, when excited by the support arm, produce directly transverse motion of the slider relative to the data track, causing head/track misregistration and, if large enough, result in servo system stability problems. While a major effort has been made within the hard disk drive industry to reduce the undesired amplification at those resonant frequencies, this effort has not been entirely successful in the past, and it has been characterized by considerable experimentation and "trial and error" techniques.
Because of the presence of the load beam flanges in the reinforced section, the shear center of the load beam is often separated from the center of gravity along the length of load beam. A pure in-plane support arm motion cart thus induce the torsional motion of the HGA even when the load beam is truly flat. Two undesirable effects result from this coupling. First, the in-plane support arm motion can excite the torsional resonance of the HGA system. This can make the head tracking system unstable if the amplification is too large, as often occurs in the hard disk drive industry. Second, the coupling of torsional stiffness with in-plane stiffness decreases the resonant frequency of the sway mode. Servo system stability problems can also arise if the sway mode amplitude and frequency are not controlled.
Heretofore, the HGA was formed to have a desired preload characteristic by forming the spring region thereof by bending it to follow a contour of a forming die. While the prior art forming dies followed e.g. a circular contour, they most frequently caused the load beam to manifest a s positive bump and offset when the HGA assumed its loaded (flying) position relative to the adjacent data storage disk surface. Attempts were made iteratively, and experimentally, to improve in-plane stiffness characteristics of the prior art HGA, but those attempts were of the "cut and try" variety, and very little scientific analysis was applied to understanding the factors and parameters affecting in-plane stiffness.
Conventional wisdom within the disk drive load beam art has suggested that ideal ,operation of the HGA in the loaded state (i.e. when the rotating disk is up to speed and the slider is flying over the disk) is realized when the load beam spring section 26 is perfectly flat, so that it is intuitively expected to manifest the greatest in-plane stiffness and rigidity. One example of this conventional wisdom is found in U.S. Pat. No. 5,065,268 to Hagen, entitled: "Load Beam Having An Additional Bend In A Head-Gimbal Assembly", the disclosure thereof being incorporated herein by reference. This prior approach featured plural transverse plastic bends defined along the spring section of the load beam in the unloaded state. When the load beam was deformed to its loaded state it was said to achieve a reasonably flat load beam to promote in-plane stiffness. If one integrates a longitudinal cross section of the spring section of that load beam, the result is approximately a "zero bump" contour. The drawbacks of this prior approach were that sharp transverse bends or creases were required during formation, leading to complex mandrel designs and fabrication procedures, and the resultant design only approximately achieved its desired goal of true maximum in-plane stiffness.
Thus, a hitherto unsolved need has remained for a load beam in the head-gimbal assembly having an optimized offset and an optimized bump to maximize loaded in-plane stiffness, and methods for making and testing the resultant load beam to achieve maximum loaded in-plane stiffness.